Image Captioning on POPE Adversarial
121.4CIDErVanilla
Evaluation Results
| Method | Links | |
|---|---|---|
| VanillaBackbone=InternVL2-7B2026.02 | 121.4 | |
| DoLaBackbone=InternVL2-7B2026.02 | 120.8 | |
| SCRBackbone=InternVL2-7B2026.02 | 120.7 | |
| OPERABackbone=InternVL2-7B2026.02 | 119.4 | |
| VLI†Backbone=InternVL2-7B2026.02 | 119.4 | |
| VanillaBackbone=Qwen2-VL-7B2026.02 | 118.5 | |
| OA-VCDBackbone=InternVL2-7B2026.02 | 118.5 | |
| SIDBackbone=InternVL2-7B2026.02 | 118.5 | |
| DoLaBackbone=Qwen2-VL-7B2026.02 | 117.9 | |
| SCRBackbone=Qwen2-VL-7B2026.02 | 117.9 | |
| Uniform-SmoothBackbone=InternVL2-7B2026.02 | 117.9 | |
| CRoPS†Backbone=InternVL2-7B2026.02 | 117.3 | |
| VCDBackbone=InternVL2-7B2026.02 | 117.2 | |
| OPERABackbone=Qwen2-VL-7B2026.02 | 116.7 | |
| VLI†Backbone=Qwen2-VL-7B2026.02 | 116.3 | |
| SIDBackbone=Qwen2-VL-7B2026.02 | 116 | |
| OA-VCDBackbone=Qwen2-VL-7B2026.02 | 115.4 | |
| VanillaBackbone=Qwen-VL2026.02 | 115.2 | |
| CRoPS†Backbone=Qwen2-VL-7B2026.02 | 115.1 | |
| Uniform-SmoothBackbone=Qwen2-VL-7B2026.02 | 115.1 | |
| SCRBackbone=Qwen-VL2026.02 | 114.6 | |
| VCDBackbone=Qwen2-VL-7B2026.02 | 114.6 | |
| DoLaBackbone=Qwen-VL2026.02 | 114.5 | |
| OPERABackbone=Qwen-VL2026.02 | 113.4 | |
| VLI†Backbone=Qwen-VL2026.02 | 113.2 | |
| OA-VCDBackbone=Qwen-VL2026.02 | 112.5 | |
| SIDBackbone=Qwen-VL2026.02 | 112.5 | |
| CRoPS†Backbone=Qwen-VL2026.02 | 111.8 | |
| Uniform-SmoothBackbone=Qwen-VL2026.02 | 111.7 | |
| VCDBackbone=Qwen-VL2026.02 | 111.2 | |
| VanillaBackbone=LLaVA-1.5-7B2026.02 | 108.5 | |
| SCRBackbone=LLaVA-1.5-7B2026.02 | 107.8 | |
| DoLaBackbone=LLaVA-1.5-7B2026.02 | 107.6 | |
| OPERABackbone=LLaVA-1.5-7B2026.02 | 106.6 | |
| VLI†Backbone=LLaVA-1.5-7B2026.02 | 106.2 | |
| SIDBackbone=LLaVA-1.5-7B2026.02 | 105.8 | |
| Uniform-SmoothBackbone=LLaVA-1.5-7B2026.02 | 105.6 | |
| OA-VCDBackbone=LLaVA-1.5-7B2026.02 | 105.3 | |
| CRoPS†Backbone=LLaVA-1.5-7B2026.02 | 104.5 | |
| VCDBackbone=LLaVA-1.5-7B2026.02 | 103.8 | |
| VanillaBackbone=Cham-7B2026.02 | 102.3 | |
| DoLaBackbone=Cham-7B2026.02 | 101.8 | |
| SCRBackbone=Cham-7B2026.02 | 101.7 | |
| VLI†Backbone=Cham-7B2026.02 | 100.6 | |
| OPERABackbone=Cham-7B2026.02 | 100.3 | |
| Uniform-SmoothBackbone=Cham-7B2026.02 | 99.8 | |
| SIDBackbone=Cham-7B2026.02 | 99.5 | |
| OA-VCDBackbone=Cham-7B2026.02 | 99.2 | |
| CRoPS†Backbone=Cham-7B2026.02 | 98.2 | |
| VCDBackbone=Cham-7B2026.02 | 97.4 |